Seismic Detectors: The State-of-the-Art

Abstract
The following detectors were reviewed: (1) IBM, (2) Z, (3) Deflection, (4) Analytic Envelope, (5) Allen, (6) Stewart, (7) Walsh, and (8) Mars. All these detectors are based on a comparison between some signal variance over a time period equal in deviation to the expected event and the normal or long time variance. Detectors using the phase have been suggested but not yet satisfactorily implemented. Uniform comparative testing of the detectors has not yet been done and is recommended. Additional research to improve detector performance is needed to: (1) develop phase sensitive detectors; (2) determine the improvements possible with matched filters; (3) further develop hybrid algorithms that combine high detection probability/high false alarm rate algorithms with low false alarm rate past processors; (4) develop algorithms designed specifically for three-component, broad-band data; and (5) investigate the use of dedicated microprocessors to run the detection algorithms on a one processor per channel (or seismic station) basis. (Author)